twoeths

from https://github.com/sigp/lighthouse/blob/v1.2.0/beacon_node/network/src/sync/range_sync/chain.rs#L682 regarding `beacon_block_by_range`, what I see is we enforce canonical blocks pretty well for archived blocks and unarchived blocks but we need to do more to make sure the...

> We could do a last loop over all blocks before sending to ensure that all blocks are part of the chain or drop them it works or `step=1` but...

> > it works or step=1 but not for step>1 > > What's the issue with step > 1 ? I mean we can't check parent-child relationship before we return...

> > I mean we can't check parent-child relationship before we return if step>1 > > Can we ask the forkchoice if blocks belong to our current head chain? forkchoice...

it's noticed that the `delaySec` is `0.88s`, and attestation of validator 270114 was not included in other AggregateAndProof as well, it's not missed because it's an aggregator in that committee...

The i/o lag only happened at that time - the below shows 4s delay between signing time and publishing time ``` May-11 05:36:27.033[] ^[[34mdebug^[[39m: Signed attestation slot=3785280, index=56, head=0xc4d2ce80c7723b88d996303a7f37ba05050f7c775277818bc689e7e6724ec000, validatorIndex=270127...

removing peerstore dir does fix the issue

libp2p failed to load a peer and it throws an error, need to wait for the latest libp2p (0.37.* atm) to see if the issue is still there ideally, libp2p...

reattaching https://github.com/ChainSafe/lodestar/issues/4603#issuecomment-1263095995 (after libp2p is updated to 0.39.2), this time it shows `restrictSize()` takes 4% Detail I disabled async stack trace so there is no async hook there @wemeetagain [0930_multilined_fastMsgId_lg1k_1.2.0.cpuprofile.zip](https://github.com/ChainSafe/lodestar/files/9694117/0930_multilined_fastMsgId_lg1k_1.2.0.cpuprofile.zip)

tracked in https://github.com/libp2p/js-libp2p-mplex/issues/205